Research on Image Segmentation Technology Based on Improved Fuzzy BP Neural Network
Research on Image Segmentation Technology Based on Improved Fuzzy BP Neural Network
复制标题
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
Zhou Chang-ying
中科院分区：
文献类型：
--
作者：
Zhou Chang-ying
针对传统的图像分割方法分辨率和清晰度较低的问题，提出了一种基于模糊BP神经网络的图像分割方法。该方法利用模糊集理论对分割图像的区域特征和特征向量进行降维处理，并根据规则确定神经元个数，输出分类决策。最后，利用模糊BP神经网络对图像进行分类。实验结果表明，该算法能有效地分割图像，分割出的图像边缘清晰，并且能有效地缩短训练时间.
This paper studies the use of BP neural network for image segmentation.Traditional image segmentation methods often cause low resolution and definition.This paper presents a fuzzy BP neural network for image segmentation.Fuzzy set theory is used to reduce the regional characteristics of segmentation images and the dimensions of feature vectors.Based on the rules,the neuron number is dicided,and the classification decision-making is output.Finally,the experiment results showthat the proposed algorithm can effectively segment images,the segmentation has sharp edges.In additon,the algorithm can shorten the training time effectiveliy.